• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 65 subscribers
  • Views 677 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Occasional delayed turn-on with TPS54615

Gregory Davis
Prodigy 70 points
Other Parts Discussed in Thread: TPS54615

                We are occasionally seeing an unexpected behavior with the turn-on of the TPS54615.  We are using it to power a XILINX FPGA. It is the last secondary voltage to be brought up on our design.

                Due to various connections to the FPGA, the 2.5V power plane is being back driven (through the FPGA) to 1.96V, prior to the TPS54615 being enabled, by other secondary voltages already up.

                Normally, we see the 1.96 volts being back driven for about 10 msecs until we enable the TPS54615, then the output climbs to 2.5V immediately and remains stable.

                However, occasionally, we see the 1.96 volts back driven even after the enable goes high for upwards to 800 msecs, then the output climbs to 2.5V and remains stable. VIN is 3.3V and is the first voltage on the board to come up.

                Can you explain why the TPS54615 doesn’t  always start up immediately after enable goes high?

                Is the back driven 1.96V causing this? 

  • 0
    TI__Guru**** 191445 points

    The TPS54615 is not designed explicitly to operate under pre-bias (or back driven) conditions.  Usually what you should see is that when EN is asserted, the pre-biased output should discharge to near 0 V before the output rises towards 2.5 V at the slow start rate.  During that discharge time, the low side FET is on constantly and the output voltage is discharged thru the low side FET.  The discharge current is controlled by the inductor.  It can ramp to exceedingly large negative values.  When teh high side FET turns on for the first time, the inductor current must remain continuous, and that current is shunted back to the input.  Sometimes this causes problems.  Usually what you will see are several attempts to discharge the pre-bias and start up.   It is possible that it may stick in this mode for a long time, then you will have to cycle power to restart successfully.  Can you post some waveforms?  Does this accurately describe what you see?

  • 0 in reply to JohnTucker
    Prodigy 70 points

    We do see the pre-biased output discharge to near 0V just prior rising to 2.5V.I will attach a picture that shows when this action occurs.

    Please disregard any small spikes on the signals as the probe was not grounded properly.

     We have never had to restart the circuit to achieve the 2.5V.  It will aways come up on its own in less than a second worse case.

    First picture is normal operation. Small interval of backdrive, then output discharge and startup.

    Second picture shows the tail end of the delayed discharge and startup.

     

  • 0 in reply to Gregory Davis
    TI__Guru**** 191445 points

    It is hard to tell form those waveforms as the time scale is 100 ms/div.  The activity of interest will all occur in the area of the trigger.  If you see a single dip towards zero, then a ramp up to Vout, then the waveform is completely normal.  If you see multiple start attempts (like a saw tooth waveform), then that is repeated re-starts probably due to excessive sink current.  It has been several years since I last worked with that device, so I do not think I have an example waveform to show you.

  • 0 in reply to JohnTucker
    Prodigy 70 points

    We do not see the sawtooth behavior.  It just appears to wait a while before trying to start up.  If you don't see any harm in the way it is behaving, we will leave it as is.

    However, if the pre-bias is a serious concern regarding start-up of this device, let us know and we'll take action to prevent it.

    Thanks...